Bug 119739 – Precision updates in WebAudio

• Component: Web Audio
• Status: RESOLVED FIXED

[Praveen Jadhav]

Attributes minDecibels, maxDecibels and smoothingTimeConstant in AnalyserNode and currentTime in AudioContext should be of type double.

[Praveen Jadhav]

Created attachment 208614 [details]
Patch

Patch includes one minor optimization in GStreamer porting file WebKitWebAudioSourceGStreamer.cpp

[Chris Dumez]

Comment on attachment 208614 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=208614&action=review

> Source/WebCore/ChangeLog:7
> +

Missing changelog

> Source/WebCore/platform/audio/gstreamer/WebKitWebAudioSourceGStreamer.cpp:-246
> -    for (unsigned channelIndex = 0; channelIndex < priv->bus->numberOfChannels(); channelIndex++) {

Looks unrelated?

[Philippe Normand]

Comment on attachment 208614 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=208614&action=review

>> Source/WebCore/platform/audio/gstreamer/WebKitWebAudioSourceGStreamer.cpp:-246
>> -    for (unsigned channelIndex = 0; channelIndex < priv->bus->numberOfChannels(); channelIndex++) {
> 
> Looks unrelated?

This is unrelated, and how is this an optimization? This is how the GLib docs advize on how to fill a GSList. It's not the first time people try to optimize this code :)

[Chris Dumez]

Comment on attachment 208614 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=208614&action=review

>>> Source/WebCore/platform/audio/gstreamer/WebKitWebAudioSourceGStreamer.cpp:-246
>>> -    for (unsigned channelIndex = 0; channelIndex < priv->bus->numberOfChannels(); channelIndex++) {
>> 
>> Looks unrelated?
> 
> This is unrelated, and how is this an optimization? This is how the GLib docs advize on how to fill a GSList. It's not the first time people try to optimize this code :)

Well, it avoids a call to g_slist_reverse() so it can only be faster IMHO. That said, this is a micro-optimization considering the number of channels (and thus the size of the list).
In any case, this should not be in this patch.

[Praveen Jadhav]

(In reply to comment #2)
> (From update of attachment 208614 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=208614&action=review
> 
> > Source/WebCore/ChangeLog:7
> > +
> 
> Missing changelog

oh, I missed. Will update this shortly.

> 
> > Source/WebCore/platform/audio/gstreamer/WebKitWebAudioSourceGStreamer.cpp:-246
> > -    for (unsigned channelIndex = 0; channelIndex < priv->bus->numberOfChannels(); channelIndex++) {
> 
> Looks unrelated?

Its a minor change to avoid calling g_slist_reverse() API. So, thought of including this in the same bug. Will raise a new bug for this patch.

[Praveen Jadhav]

Created attachment 208616 [details]
Patch

Patch updated as per the comments.

[WebKit Commit Bot]

Comment on attachment 208616 [details]
Patch

Clearing flags on attachment: 208616

Committed r155112: <http://trac.webkit.org/changeset/155112>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.